Arakawa and colleagues (2011) use temporal changes in obsidian source patterns to link the late thirteenth-century abandonment of the Mesa Verde region to Ortman's (2010, 2012) model of Tewa migration to the northern Rio Grande. They employ Anthony's (1990) concept of reverse migration, inferring that an increase in Mesa Verde–region obsidian from a specific Jemez Mountain source reflects the scouting of an eventual migration path. Weaknesses of this inference are that only obsidian data from the Mesa Verde region were used in its development and that the model does not consider the complexities of previously documented patterns of settlement and stone raw material use in the northern Rio Grande. By examining source data from parts of northwestern and north-central New Mexico, we find that the patterning seen in the Mesa Verde obsidian data is widespread both geographically and temporally. The patterns are more indicative of a change in acquisition within a down-the-line exchange system than a reverse migration stream. Population trends on the southern Pajarito Plateau, the probable source of the acquisition change, suggest ancestral Keres rather than Tewa involvement in thirteenth-century obsidian distribution.